In a tuner or frequency converter (such as, for example, a dual-conversion tuner), an incoming signal at frequency fIN is mixed with a signal at frequency fLO1 from a local oscillator (LO) to produce a signal at an intermediate frequency fIF. This signal may then mixed with a signal at frequency fLO2 from a second local oscillator signal to produce the desired output frequency fOUT, in a dual conversion tuner configuration. This process is illustrated in FIG. 1A, which is a portion of one example of a tuner showing how the fLO signals (provided by LO 12 and 13) are mixed. Such a tuner is shown in U.S. Pat. No. 5,737,035, issued Apr. 7, 1998 hereby incorporated by reference herein. Typically, but not always, the frequency of first local oscillator, e.g., LO 12, is greater than that of second local oscillator, e.g., LO 13. That is, generally fLO1>fLO2. Accordingly, reference shall be made herein to equations in which it is assumed that fLO1>fLO2. However, it should be appreciated that the formulae herein are applicable to situations in which fLO2>fLO1, such as by replacing fLO1 with fLO2 and replacing fLO2 with fLO1 in situations where fLO2>fLO1.
FIG. 1B shows a simplified diagram of two mixing stages with the filtering omitted. These filters ultimately determine final bandwidth (fBW) of the tuner, but since they do not contribute to the production of LO-related spurs, they are omitted from FIG. 1B.
An adverse effect of the dual conversion process is the introduction of LO-related spurs into the tuned signal. These spurs are created by combinations of the harmonics of the LO frequencies used (fLO1 and fLO2).
The frequency of each of the LO-related spurs can be calculated as:fSPUR=n×f1−m×f2  (1)where n and m are integer numbers representing, respectively the harmonics of the high and low local oscillator frequencies, and f1 and f2 are the local oscillator frequencies (e.g., fLO1 and fLO2, respectively where fLO1>fLO2). If any spur generated by a given combination of fLO1 and fLO2 falls within the output bandwidth (fBW) of the converter/tuner, that spur can degrade the quality of the output signal. If a spur does exist within the desired output bandwidth, the LO frequencies can be adjusted to different values to avoid the spur falling within the output band. As manufacturing processes produce denser and faster IC's, the number of harmonics (nMAX) that must be considered continues to increase. Since the number of LO frequency combinations that can possibly create spurs in n harmonics is n2, the amount of resources required to avoid the spurs increases dramatically as technology improves. As an example, at the time the circuit shown in FIG. 1A was initially produced, the number of harmonics (n) that were typically taken into consideration was 5. Currently, the number of harmonics typically taken into consideration is on the order of 15.
One reason why it is important to avoid LO spurious products is that a spur which is generated by multiples of fLO1 and fLO2 in a double conversion system will often have a power level which is much greater than the actual RF signal. Therefore, if a spur caused by a product of fLO1 and fLO2 falls in the desired IF output pass band, its amplitude (power level) may be larger than the IF output level of the original desired signal, corrupting the performance of the mixer itself.
One of the fixes for this problem is that when it is known that a certain spur (such as a spur associated with two times the first LO and three times the second LO) will fall within the output pass band, the LO frequencies can be changed (up or down) a certain amount, which will, in effect, still allow the circuit to tune to the desired output frequency, but the spur will be moved up or down and outside of the output bandwidth of the tuner.
Accordingly, one method for identifying spurs falling within a particular band, such as the tuner output band, is to look at all the harmonics of the first LO, mixed with all the harmonics of the second LO and, one by one, check off each one. Thus, if a circuit designer is looking up to the 15th harmonic of the first LO and the 15th harmonic of the second LO, the designer checks one times fLO1 (first harmonic) and one times fLO2 (first harmonic) to see if there is a spur of concern. If there is no spur of concern, then the designer continues with one times fLO1 (first harmonic) and two times fLO2 (second harmonic) to see if there is a spur of concern. If not, then the process continues with one times fLO1 (first harmonic) and three times fLO2 (third harmonic) to see if there is a spur of concern. Once all harmonics of fLO2 have been considered, the harmonic of the first LO frequency may be incremented and each harmonic of the second LO frequency again considered. That is, the designer continues with two times fLO1 (second harmonic) and one times fLO2 (first harmonic) to see if there is a spur of concern, and so on. This results in n2 combinations being looked at. This is a time consuming method. Even assuming that the mathematics of how spurs are generated allows for the elimination of quite a few of the coefficients for the first and second LO, the operation remains essentially an n2 operation.
It should be appreciated that spur identification and avoidance as discussed above is dependent on the circuit that is being used and which spurs might come through the chip more strongly than other spurs. It is also dependent on the input frequency and on all the specific channels that might be on the input frequency. That method is also specific to the first IF frequency and to the output frequency. Thus, for each application of a circuit the chip designer generally must employ a unique program for each channel input lineup in the desired frequency spectrum. This then implies that a different spur avoidance algorithm must be created for every customer application, i.e., each tuner implementation.